From AI-generated content to democratic resilience: practical media literacy with young people

Introduction

Generative AI has made the information environment faster, cheaper to manipulate and harder to read at a glance, but the educational challenge is broader than learning to recognise synthetic content. Young people need practical ways to understand how information reaches them, test claims against evidence and decide what deserves their confidence. We propose a form of media literacy that moves from detection to judgement, and from judgement to responsible participation.

Our central concern is democratic resilience. In a contested information environment, resilience does not mean distrusting everything; it means being able to investigate, tolerate uncertainty, revise conclusions and continue participating in public life. The following sections examine why conventional 'spot the fake' activities are insufficient, how participatory methods can turn media literacy into a lived practice, and what youth organisations and educators can do to make those capabilities transferable beyond a single workshop.

How generative AI is changing the information environment

Generative AI increasingly makes the appearance of informational authority easier to manufacture, and even easier to share. Fluent explanations, convincing images and synthetic recordings can be produced quickly and adapted to different audiences. Misleading claims can therefore acquire the presentation of credible information without the research or accountability that credibility should require. 

Recognising AI-generated content addresses only part of this problem. A synthetic image may be clearly labelled and used responsibly; an authentic photograph may be attached to a false claim. Similarly, an AI-generated explanation can contain accurate information alongside unsupported assertions. Its confident wording gives the reader little basis for distinguishing between them (Miao and Holmes, 2023). We argue that educational responses should make the relationship between a claim and its evidence the central question.

The democratic consequences also extend beyond accepting falsehoods. Political deepfakes can generate uncertainty and weaken trust in news, even when audiences are not persuaded by their claims (Vaccari and Chadwick, 2020). An approach that repeatedly demonstrates how easily information can be fabricated, without teaching how to investigate it, risks leaving young people doubtful but poorly equipped to judge.

Media literacy should give that judgement practical foundations. Young people need to understand what generative tools can produce, how claims reach them and what would justify confidence in those claims. This supports their ability to use AI constructively while retaining responsibility for what they believe and communicate. We see this as one element of democratic resilience: enabling young people to assess public claims and remain willing to participate, even when information is contested.

The limitations of conventional “spot the fake” approaches

“Spot the fake” exercises can make misinformation tangible, but their value depends on what learners can do afterwards. Recognising a manipulated image or fabricated headline demonstrates success on a particular task. It does not establish that someone can investigate an unfamiliar claim, assess conflicting evidence or make a responsible sharing decision. We argue that media-literacy education should develop these transferable capabilities through sustained practice.

Exercises focused on surface clues can leave learners examining a post without investigating its source. Establishing credibility often requires checking beyond the original content. Young people need to establish who is making a claim, trace its supporting evidence and check whether other accounts offer independent corroboration. Several websites repeating the same unsupported allegation do not equal better evidence. Lateral reading, which involves leaving the original page to investigate its source and claims elsewhere, offers a practical foundation for this approach (Wineburg and McGrew, 2019).

Moreover, evaluation requires contextual judgement. Information can mislead through selective presentation, omitted context, or an interpretation that exceeds what the evidence can reasonably prove. Evaluating a claim requires examining who produced it, how it has been presented and what its supporting evidence actually shows (Wardle and Derakhshan, 2017). For education, this means helping learners explain what a source establishes and where its limits lie. Public debate contains disagreements over policy and values as well as disputes over facts; treating them all as questions of “real” or “fake” obscures distinctions essential to democratic discussion.

Knowledge must also inform behaviour. The ability to verify a claim has limited protective value if accuracy is overlooked when deciding to share it. Evidence that directing attention to accuracy can improve sharing decisions supports making this moment of action part of media-literacy education (Pennycook et al., 2021). Young people should practise checking claims they find appealing, communicating uncertainty and correcting their own contributions when the evidence changes.

We therefore advocate an approach whose success is judged by how learners reason and participate beyond the exercise. Can they investigate a source, justify their confidence and revise a conclusion? Can they scrutinise powerful actors while recognising credible evidence? These capabilities give young people grounds for informed participation. This approach also builds on the policy report Navigating the Information Disorder, which advocated media literacy co-designed with young people and proposed peer-led reflection and narrative framing exercises for schools and youth organisations (Europinion, 2025). These formats offer a practical starting point for helping young people examine why a claim seems credible and how its presentation shapes interpretation. Their value should be judged by whether participants apply that judgement beyond the workshop (Europinion, 2025).

Participatory media literacy and democratic resilience among young people

Disinformation is often a symptom of a wider problem: how information is consumed. Young people do not encounter claims in a neutral space. They encounter them inside feeds shaped by algorithms, repetition, social signals and emotionally engaging content. Contemporary media education therefore needs to address the role of algorithms alongside source analysis and other established media literacy skills (OECD, 2024). Media literacy should help young people ask not only 'is this true?' but also 'why am I seeing this?', 'what is trying to hold my attention?', 'who benefits from this framing?' and 'what would make me change my mind?'.

Participation is an imperative to instilling media literacy and democratic resilience amongst young people. The youth of today should not be treated as passive recipients of warnings about online risks, but rather as partners for they can bring the examples, habits and tensions of their own information environments into the learning process and examine them together. For instance, facilitated face-to-face work can create a space to compare interpretations, trace evidence, discuss why certain narratives feel persuasive and practise disagreement without collapsing into either certainty or cynicism. This participatory approach is consistent with European policy that explicitly identifies non-formal and informal learning and youth work as settings in which media literacy, critical thinking and resilience to manipulated information can be strengthened (European Commission, 2023). Democratic resilience depends on the capacity to remain curious, critical and willing to engage with others across disagreement.

Participatory media literacy also allows education to address the social and emotional dimensions of information consumption. For instance, echo chambers are both sustained by a lack of facts as well as a desire for belonging and identity. Research has demonstrated that moral and emotional language can increase the diffusion of political messages within ideological networks, as they are designed specifically to resonate to one’s emotions rather than one’s reason (Brady et al., 2017). The reward structures of digital platforms can further shape what people notice and share. Activities should therefore make room for reflection on attention and emotion alongside verification. Ongoing cooperation on disinformation with partners in New Caledonia, for example, strengthens the case for approaches that connect information literacy with lived political and community contexts rather than presenting misinformation as an abstract digital problem.

Finally, participation should not be confused with activity for its own sake. A workshop is not successful simply because participants enjoyed it or correctly identified a set of fabricated examples. The more meaningful question is whether their reasoning changes. Evidence suggests that media-literacy learning can improve young people's judgments of evidence and that even relatively short digital media-literacy interventions can improve discernment between higher- and lower-quality information (Kahne and Bowyer, 2017; Guess et al., 2020). Can participants explain why they trust a source? Do they seek corroboration? Do they pause before sharing? Can they recognise when evidence is insufficient and communicate uncertainty? Pre- and post-intervention surveys can be combined with facilitated reflection and follow-up tasks to assess these changes. A facilitator resource bank can then help youth workers reproduce the approach while adapting examples to the communities they know best.

Practical recommendations for youth organisations and educators

1. Start with the information journey, not only the artefact. Ask who created a claim, where it originated, why it reached the learner and what incentives or platform dynamics may be shaping its visibility. AI-generated content should be treated as one part of that wider environment. Understanding algorithms should be treated as an integral part of contemporary media education rather than an optional extra (OECD, 2024).

2. Practise verification as a repeatable behaviour. Use unfamiliar, realistic claims and require learners to leave the original post, investigate the source, trace evidence and look for independent corroboration. The goal is a method they can use later, not success on a prepared quiz.

3. Make algorithms and attention discussable. Invite participants to reflect on why particular content appears in their feeds, how repetition and engagement signals can shape perceived credibility, and how emotion affects what they click, believe and share.

4. Co-design with young people and use peer learning. Build activities around examples that participants recognise from their own information environments. Give them roles as investigators, explainers and facilitators so that media literacy becomes something they practise with others rather than something delivered to them. This also reflects the European Commission's emphasis on meaningful youth participation and on strengthening media literacy through youth work and non-formal learning (European Commission, 2023).

5. Connect judgement to action. Include the moment of sharing in every exercise. Learners should practise pausing, communicating uncertainty, choosing not to amplify weak claims and correcting their own contributions when better evidence appears.

6. Measure transfer, not just attendance. Use pre- and post-intervention questions, reflection prompts and follow-up exercises to assess whether participants have changed how they investigate and justify claims. Useful indicators should capture confidence in a conclusion, evidence-seeking behaviour and responsible sharing decisions, rather than only whether an activity was completed. Evaluating these behavioural and judgement outcomes matters because media-literacy interventions can produce measurable improvements in discernment that extend beyond performance on a single exercise (Guess et al., 2020).

7. Protect trust as well as scepticism. Media literacy should not leave young people believing that nothing can be known. Educators should model what credible evidence looks like, explain the limits of a source and show that changing one's mind in response to better evidence is a strength. The objective is informed participation, not permanent suspicion.

The purpose is not to turn every young person into a professional fact-checker. It is to make reflective verification, responsible sharing and evidence-based disagreement normal parts of participation. In an information environment increasingly shaped by generative AI and algorithmic distribution, those habits are not only media-literacy skills; they are democratic skills. European policy increasingly makes the same connection between media literacy, resilience to disinformation and effective democratic participation (European Commission, 2023).

Image: Will Allen/Europinion                                                             

All rights reserved.